Atmospheric Environment 37 (2003) 4425–4434

A comparison of PM10 monitors at a Kerbside sitein the northeast of England

Monica Pricea,*, Susan Bulpitta, Michael B. Meyerb

aSchool of Health, Natural & Social Sciences, University of Sunderland, Sunderland, UKbRupprecht & Patashnick Co., Inc., Albany, NY, USA

Received 16 December 2002; received in revised form 26 June 2003; accepted 3 July 2003

Abstract

There is a need for a consistent measurement technique for both PM10 and PM2.5 that is capable of providing real-

time data suitable for determining the effects of particulate pollution on human health. Rupprecht and Patashnik have

developed a TEOMs monitor configuration that increases collection efficiency for the semi-volatile mass fraction, when

present. By operating at a lower setpoint temperature the system offers a real-time monitor that removes particle bound

moisture and promises to improve comparability with the European Union (EU) reference gravimetric method. Trials

with the device, a conventionally operated TEOM and a Partisols gravimetric monitor have shown that in the

northeast of England the loss of organics and nitrates may not be the major cause of the observed differences between

the monitors. Instead the data presented in this study indicate that it is the retention of particle bound water by the EU

reference method that may be causing the observed differences.

The presence and amounts of moisture associated with particles depends on the chemical composition and size range

of the particles as well as the ambient relative humidity. As both of these factors vary spatially and temporally it is

problematic to apply scaling factors to make data collected by the TEOM comparable to data collected by the EU

reference method. In addition, whether particle bound moisture, some of which may be absorbed after sampling should

be included in air quality standards needs further investigation.

r 2003 Elsevier Ltd. All rights reserved.

Keywords: Atmospheric particles; Measurement; Moisture; Semi-volatile matter

1. Introduction

Within the European Union (EU) the specified

reference method for monitoring PM10, EN12341, is a

gravimetric technique which requires that the PM10

mass fraction is collected upon a filter. Filters are

normally exposed for 24 h; although this can be

extended if the concentration of particulate matter is

thought to be low. Preparation of filters for exposure

must be carried out under strict, specified conditions.

Prior to, and after exposure, equilibration must take

place at 2071�C and a relative humidity of 5075% for

24 h (unless further equilibration is indicated). Filters

have to be weighed using a balance accurate to within

10 mg (BSI, 1998). Under the terms of the first daughter

directive on air quality, 1999/30/EC, EU member states

must monitor concentrations of atmospheric PM10 using

methods that demonstrate equivalence to the reference

method. Member states may use alternative techniques

providing they can illustrate a consistent relationship to

the reference method. Any results obtained must be

adjusted by a relevant factor to show equivalence with

the results that would have been obtained, should the

reference method have been used.

The requirement for filter equilibration and the nature

of the operation of gravimetric monitors means that

reference method results are often not obtained for a

ARTICLE IN PRESS

AE International – Europe

*Corresponding author.

E-mail address: monica.price@sunderland.ac.uk (M. Price).

1352-2310/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.

doi:10.1016/S1352-2310(03)00582-X

period of up to 2 weeks. However, in terms of the

determination of health effects and the issue of public

information, there is a need for real-time data. In order

to meet this requirement many countries are adopting

the tapered element oscillating microbalance (TEOM

Monitor), which has been in use world-wide since 1988,

to give an automatic, near real time measurement of

particulate matter (Meyer et al., 2000).

The TEOM method is based on the simple physical

law that the frequency of mechanical oscillation of an

element is directly proportional to the mass of an

element. The TEOM method uses a standard regulatory

PM10 inlet, operating at 16:7 l min�1; to select the

particles. These then pass through an isokinetic

flow splitter from which 3 l min�1 passes through the

16 mm diameter filter cartridge (13:5mm active area

sampling diameter) connected to the top of a narrow

oscillating hollow tapered glass tube. The collection

of particles by the filter at the free end of the

tapered glass tube will alter its effective mass, which in

turn will change its resonant frequency. As more

particles deposit on the filter the tube’s natural

frequency of oscillation decreases (QUARG, 1996). A

microprocessor converts the oscillation frequency to

mass and then to mass concentrations, which are

updated every 2 s: The inlet and the sensing system of

the TEOM are operated at a constant temperature of

50�C in order to minimise the adsorption/desorption

effects of atmospheric moisture on the microbalance

filter (Mukerjee et al., 1999).

Concerns have arisen about the TEOM monitor due

to the default operating temperature of 50�C: It is

believed that in addition to the removal of water vapour,

organic compounds and a proportion of the sulphate

and (more likely) nitrate may also be lost. This has

resulted in the TEOM monitor sometimes appearing to

give lower concentrations than the reference method in

Northern Europe (Muir, 2000).

A number of studies have presented data concerning

the loss of semi-volatile species on the filter of the

TEOM relative to the manual gravimetric method (Allen

et al., 1997; Ayers et al., 1999; Meyer et al., 1992).

APEG (1999) concluded that in the UK at concentra-

tions around 50 mg m�3 the TEOM tends to under-read

compared with a gravimetric sampler by between 15%

and 30%. In areas where semi-volatile components

constitute a high fraction of PM10 it would be desirable

to maintain an operating temperature as low as possible,

or as close to the reference method equilibration

temperature, to minimise the loss of these semi-volatile

fractions.

In order to provide a real-time monitor that compares

more readily with the EU reference method, Rupprecht

and Patashnick have developed a system that enables the

sample air stream flowing into a TEOM monitor to be

operated at a lower temperature. The sample equilibra-

tion system (SES) continuously conditions the flows of

the TEOM monitor to remove moisture, allowing the

instrument to be operated year-round at a temperature

closer to that of the gravimetric reference method

(Meyer et al., 2000). A reduction in the operating

temperature of the TEOM monitor should allow the

greater retention of semi-volatile material.

In order to remove moisture the SES incorporates a

Nafions dryer, which continuously conditions both the

sample and bypass flows of the TEOM monitor. The

SES enables filter-based mass measurements by the

TEOMmonitor under consistent, dry conditions (Meyer

et al., 2000) and at a reduced temperature. In a

validation study using a monitor fitted with a Nafion

drier Eatough et al. (2003) have demonstrated that fine

particles are not lost during the passage of air through

the drier system.

Initial trials with a TEOM monitor fitted with

an SES, to monitor PM2.5, were carried out in

Albany, New York, USA (Meyer et al., 2000). In this

study it was found that the TEOM monitor plus SES,

operated at 30�C; consistently recorded a higher

concentration of particulate matter than the TEOM

monitor operating at 50�C: The SES TEOM monitor

also showed a better relationship to a gravimetric

monitor than the conventional TEOM monitor

configuration.

One of the first sites in the EU to monitor particulate

matter utilising an SES modified TEOM was Trimdon

Street, Sunderland, northeast England. At a kerbside

monitoring station, PM10 has been monitored using

three machines: a TEOM Series1400a monitor operating

at 50�C; a Partisol—Plus model 2025 air sampler, and a

TEOM series 1400a PM monitor fitted with an SES,

operating at 30�C: This paper reports the findings of a

trial conducted between November 2000 and August

2001.

2. Methods

2.1. Sites

All three monitors were sited at the City of Sunder-

land kerside monitoring unit (sited 1m from the kerb) in

the northeast of England. The site is situated adjacent to

a busy road junction with traffic flows of

2000 vehicles h�1: Traffic flows are consistently high

throughout the day. Fig. 1 illustrates the location of

Sunderland within the northeast of England and the

location of the sampling site within the City.

2.2. Monitors

Monitoring for PM10 was carried out using a Partisol

Plus 2025 air sampler, a tapered element oscillating

ARTICLE IN PRESSM. Price et al. / Atmospheric Environment 37 (2003) 4425–44344426

microbalance TEOM series 1400 operating at 50�C

and a TEOM monitor plus SES drier operating

at 30�C; all manufactured by Rupprecht &

Pataschnik Co. The addition of a drier to the

TEOM device has been described elsewhere (Meyer

et al., 2000). Monitoring using the co-located

devices commenced at the site in November 2000.

Total nitrogen oxides, nitric oxide and nitrogen

dioxide are also monitored at the site by Sunderland

city council.

Partisol filters were equilibrated pre- and post-

sampling according to the requirements of EN12341

and weighed using a balance accurate to 0:001 mg; withreplicate weighings of each filter being made. Initial

investigations of the method had shown that there was

no further weight loss from filters after the recom-

mended equilibration period. Filters were exposed in the

monitor for a period of 24 h:Flow rates through both TEOM monitors are

16:7 l min�1 with 3 l min�1 passing through the sampling

head to a 13:5 mm filter. The Partisol has a flow rate of

16:7 l min�1 passing through a filter of 47 mm diameter

(active area of 42 mm), thus the filter face velocities of

the air passing through the filters are comparable.

Particulate matter was collected onto PTFE-coated glass

fibre filters (Pall Corporation).

Meteorological data for the sampling period was

obtained from the University meteorological station in

Sunderland.

3. Results and discussion

3.1. Analysis of PM10 data

The PM10 data collected by the three monitors

appears in Figs. 2–4 and the correlation coefficients in

Table 1. Good correlation is displayed between the

TEOM and the TEOM plus SES all year round.

However, the correlation between these monitors and

the Partisol sampler appears to be seasonal, with the

strongest correlation occurring during the summer

months. Figs. 2–4 display the time series for each

sampler (where the TEOM and TEOM plus SES are

averaged over 24 h periods to match the sampling period

of the Partisol). During periods of elevated PM10 in the

winter and spring, the Partisol sampler records higher

mass than the TEOM and the TEOM plus SES.

However, during the summer all three instruments tend

to record elevated PM10 concentrations.

The wintertime recorded differences between the

Partisol and TEOM monitors occurred during episodes

of elevated PM10, with the gravimetric monitor record-

ing a greater increase than the TEOM monitor. The

wintertime discrepancy in PM10 concentrations as

measured by the Partisol and TEOM has been observed

in previous studies such as that by Green et al. (2000).

This seasonal difference in the relationship is shown in

the strength of correlations (Table 1). In November the

correlation coefficient for the relationship between the

ARTICLE IN PRESS

0 5kms

Scale

0 0.5km

Scale

N

N

Morpeth

Biyth

A19

A69Newcastle

Chester-le-Street

Durham

Gateshead

Area ofdetail map

ENGLAND

A1(

M)

A19

Middlesborough

Hartlepool

Washington

A1(

T)

South Shields

SUNDERLAND

Weatherstation

Monitoring site

City centre

River Wear

NorthSea

Fig. 1. Location of Sunderland in the northeast of England and the site within the city of Sunderland.

M. Price et al. / Atmospheric Environment 37 (2003) 4425–4434 4427

Partisol monitor and the TEOMmonitor is 0.59 whereas

in August it is 0.89. Throughout the time period the

results obtained for the TEOM and the TEOM plus SES

show a strong relationship.

Recorded differences between the TEOM and gravi-

metric monitors are well documented. They include the

work of Salter and Parsons (1999), who used a TEOM

1400a series monitor operating at 3 l min�1 airflow and a

Partisol Model 2000 air sampler operating at

16:7 l min�1: In this study the TEOM was operated at

50�C and results were collected for PM10 over a period

of 100 days from April to October 1997. The results

obtained showed that although the TEOM and Partisol

results correlate well at low values of PM10 as particulate

ARTICLE IN PRESS

0

20

40

60

80

100

120

140

1/3 5/3 9/3 13/3 17/3 21/3 25/3 29/3 2/4 6/4 10/4 14/4 18/4 22/4 26/4 30/4 4/5 8/5 12/5 16/5 20/5 24/5 28/5

date

PM

10 µg

m-3

Partisol TEOM TEOM+SES

Fig. 3. PM10 ðmg m�3Þ results for March 2001–May 2001.

Fig. 2. PM10 ðmgm�3Þ results for November 2000–January 2001.

M. Price et al. / Atmospheric Environment 37 (2003) 4425–44344428

concentrations increased the Partisol was found to

record a higher concentration than the TEOM. A study

by Cyrys et al. (2001) evaluated the difference between a

TEOM 1400a series monitor operated at 50�C and air

flow of 3 l min�1 with a Harvard–Marple impactor,

using Teflon membrane filters (PTFE). Data were

collected for the period March–April 1999 and the

TEOM consistently recorded lower values than the

Harvard Impactor.

Both of the above studies collected data over a

relatively short time period even though the composition

of particulate matter is reported to vary on a seasonal

basis (Harrison et al., 1997). The differing composition

of sampled PM10 may lead to both spatial and temporal

differences in the recorded discrepancies between

monitors. The results collected in this study illustrate

the seasonal differences and highlight the necessity for

monitoring over an extended time scale.

In terms of the compositional differences between

PM10, as sampled by gravimetric monitors and the

TEOM monitor operated at 50�C; a number of studies,

including those by Green et al. (2001), Chung et al.

(2001) and Allen et al. (1997) suggest that the elevated

temperature of the TEOM monitor causes the loss of

semi-volatile material including sulphate and nitrate.

The results from the present study would suggest that

this is not the case at the Sunderland site. The TEOM

monitor plus SES drier operated at 30�C; records a

comparable value to the TEOM monitor operated at

50�C; and for some periods of the study this was lower.

These results contrast to those reported byMeyer et al.

(2000) in an initial trial with the SES carried out at

Albany, New York, USA during the summer. During

the Albany study temperatures were high, average

daytime temperature was 30�C dropping to 10–15�C

ARTICLE IN PRESS

0

10

20

30

40

50

60

70

80

90

100

1/6 7/6 13/6 19/6 25/6 1/7 7/7 13/7 19/7 25/7 31/7 6/8 12/8 18/8 24/8 30/8

date

Partisol TEOM TEOM+SES

PM

10 µg

m-3

Fig. 4. PM10 ðmgm�3Þ results for June 2001–August 2001.

Table 1

Correlation coefficients for the different monitors

TEOM SES

Full data set Partisol 0.675a 0.743a

TEOM — 0.867a

November Partisol 0.589a 0.502a

TEOM — 0.879a

December Partisol 0.746a 0.680a

TEOM — 0.952a

January Partisol 0.403a 0.608a

TEOM — 0.867a

March Partisol 0.668a 0.664a

TEOM — 0.931a

April Partisol — 0.650a

TEOM — 0.936a

May Partisol 0.737a 0.845a

TEOM — 0.918a

June Partisol 0.794a 0.841a

TEOM — 0.932a

July Partisol — —

TEOM — —

August Partisol 0.887a 0.854a

TEOM — 0.975a

aSignificant correlation at pp0:05:

M. Price et al. / Atmospheric Environment 37 (2003) 4425–4434 4429

at night. The recorded relative humidity was high at

night reaching 95–100% but lower during the daytime

dropping to 40–50%. The results obtained by Meyer

et al. showed an increased quantity of PM2.5 monitored

by the TEOM plus SES monitor as compared to the

conventional TEOM monitor. The results obtained by

the TEOM plus SES being comparable to those

obtained by a gravimetric monitor.

At the Sunderland site, weather conditions are

different to those in Albany, during the winter

temperatures range from nighttime lows of –5–8�C

and daytime temperatures of 5–12�C. Relative humidity

is high throughout the day and night (80–100%). In the

summer, temperatures are found to range from night-

time lows of 6–20�C, to daytime temperatures of 12–

25�C. Relative humidity in the summer is lower ranging

from 55–85%. At this site both TEOM monitors

correlate well with the gravimetric monitor during the

summer but not during the winter. The contrasting

results indicate that the difference between PM10

monitors will be dependent upon a range of factors,

including meteorology. The results from one study

should therefore not be applied to another area of

contrasting meteorology and particulate composition.

In the winter at the Sunderland site, the TEOM plus

SES and the TEOM operating at 50�C show a good

correlation, suggesting that the loss of semi-volatile

materials may not be occurring at this site.

In order to obtain an accurate measure of atmo-

spheric particulate matter, the retention of semi-volatile

material will be of importance if the composition of the

atmospheric aerosol comprises materials that will be lost

at 50�C, and under certain circumstances this may

comprise materials other than nitrate. Moisture could

also be considered as a semi-volatile material lost by the

50�C TEOM monitor. The default operating tempera-

ture of the TEOM monitor is chosen to minimise the

amounts of particle bound water monitored whereas

gravimetric monitors rely on filter equilibration to

remove moisture.

A study carried out in Canada (Mignacca and Stubbs,

1999) showed that a reduction in the operating

temperature of the TEOM from 50�C to 30�C resulted

in a 22% increase in measured PM10. However, it is

unclear whether this increase is due to the retention of

particle bound moisture or semi-volatile nitrate and

organics. Without a full knowledge of the composition

of the atmospheric aerosol it is not possible to determine

the nature of the differences recorded in terms of the

composition of the material lost. In some areas the

material lost could be organics or nitrate, in other areas

it may be particle bound moisture.

Atmospheric aerosol is a complex mixture of inor-

ganic and organic components where organic species can

represent up to 50% of the aerosol mass depending on

location (Cruz and Pandis, 2000). Whilst the water

uptake of the inorganic fraction has been widely

investigated and is well understood the knowledge

gained concerning the organic fraction has been limited.

In terms of the inorganic fraction at a low relative

humidity aerosol particles comprising inorganic salts are

solid (Seinfeld and Pandis, 1998). However, as the

relative humidity increases a threshold value will be

reached above which water will be rapidly absorbed by

the particle. The relative humidity at which this occurs is

characteristic for the aerosol composition and is termed

the deliquescence relative humidity. Values for typical

components of atmospheric aerosols are: NaCl—75.3%,

NH4NO3—61.8%, (NH4)2SO4—79.9% (Seinfeld and

Pandis, 1998). These are all values of relative humidity

exceeded in Sunderland routinely in the winter but also

during the spring and summer. Seinfeld and Pandis

(1998) note that not all aerosol species exhibit deliques-

cence behaviour, some such as H2SO4 are hygroscopic

and the water content associated with them changes

smoothly as the relative humidity increases or decreases.

Studies with the organic fraction have shown that a

significant fraction of the particulate organic carbon in

the atmosphere is water soluble organic carbon (Yang

et al., 2003) which due to its affinity with water plays a

role in aerosol–cloud interactions, wet scavenging and

the formation of atmospheric haze. The affinity of the

material for water should also perhaps be considered in

terms of particle bound moisture and the measurement

of PM10.

Whilst the effects of moisture on the inorganic and

organic fraction has been considered the effects of

moisture absorption by internal mixtures of salts and

insoluble components, for example soot, is more difficult

to predict (Ebert et al., 2002). The authors report that

organic components can alter the hygroscopic behaviour

of salt particles severely.

The quantity of water associated with PM10 will

therefore be dependent on its composition and the

relative humidity of the atmosphere. This is likely to

vary both seasonally and temporally and with local and

regional pollution sources. The role of particle bound

moisture in mixed aerosols is an area that requires

further research not only in the atmosphere but also by

particles collected on the filters of particulate monitors.

Whilst the TEOM is operated at 50�C in order to

minimise particle bound water, equilibration of Partisol

filters according to EN12341(BSI 1998) is carried out to

minimise the loss of volatile material which may include

particle bound moisture. The results obtained in the

present study suggest that any water associated with the

PM10 collected on the filters may also be retained. The

correlation between the 30�C TEOM plus SES drier and

the TEOM operated at 50�C and the observed

differences to the Partisol, most noticeably in the winter,

would suggest that moisture plays a role in the

differences between monitors at the Sunderland site.

ARTICLE IN PRESSM. Price et al. / Atmospheric Environment 37 (2003) 4425–44344430

As noted previously the deliquescent relative humidity

for particulate matter commonly found in the atmo-

sphere is of the order of 70% RH. However, although

the particles will absorb moisture at a relative humidity

of 70% they do not start to release this moisture until a

much lower RH is achieved, the hysteresis phenomenon.

For (NH4)2SO4 this may not occur until the relative

humidity is reduced to 30–40% (Seinfeld and Pandis,

1998) hence the equilibration relative humidity chosen

for Partisol filters (50%) may not be sufficiently reduced

to ensure moisture is removed from the particulate mass.

If this is the case the differences between gravimetric

monitors and real-time monitors, operated to minimise

moisture, will be dependent on the amount of moisture

associated with the particulate matter collected on the

filter of the gravimetric monitor. This in turn will be

determined by the composition of the aerosol and

atmospheric relative humidity.

3.2. Relationship to meteorological parameters

In order to review the role of particle bound water

further, the data have been compared to the meteor-

ological data collected for the same period. Tables 2 and

3 present data for the meteorological conditions at times

when the greatest discrepancy ð> 20 mg m�3Þ is found

between the monitors and also at times of good

agreement (o5 mg m�3). The meteorological data were

averaged for the 24 h period of interest to allow

comparison with the PM10 data.

Rainfall was recorded on several of the days when

high discrepancies between the Partisol and TEOM and

TEOM plus SES monitors were observed. In addition

high NOx concentrations have been recorded on all of

the dates with the exception of 9/6, suggesting that both

pollution levels, including a high proportion of fresh

aerosol and atmospheric moisture have been high at the

same time. The effect of the timing of a rainfall event on

the particle bound moisture content of aerosols collected

on filters remains unclear. For example, if rainfall occurs

before sampling, the particles may be washed out of the

atmosphere thus reducing their concentration. However,

if rainfall occurs after the filter sample has been collected

by the gravimetric monitor, the movement of a moist air

stream over the collected particles could lead to the

absorption of moisture on both particles and the filter.

Of the remaining dates other than 21/12 the dew point

had been reached indicating condensation of atmo-

spheric moisture will be occurring.

The timing of increased atmospheric moisture content

and the presence of high particle concentrations may

explain the contrasting results obtained in the initial

study carried out with the TEOM plus SES (Meyer et al.,

2000) and the results obtained in Sunderland. In the

initial trials relative humidity was high at night and

reduced during the day whereas in Sunderland relative

ARTICLE IN PRESS

Table

2

Datesofgreatest

discrepancy

(>20mg

m�3)betweenthePartisolandTEOM

monitors

Date

Partisol

TEOM

TEOM+SES

NO

xMin

temp

Maxtemp

Dew

point

RH

Rainfall

Rainfall

Sunshine

Meanwindspeed

Winddirection

(mgm

�3)

(mgm

�3)

(mgm

�3)

(ppb)

(�C)

(�C)

(�C)

(%)

(mm)

(h)

(h)

(ms�

1)

7/12/00

71.5

40.9

39.3

108

6.4

10.0

695

16.0

7.2

0.6

3.8

SW

11/12/00

75.2

55.8

45.0

94

6.8

12.6

986

00

2.4

3.8

SW

18/12/00

82.4

32.2

33.1

147

�0.5

5.4

098

00

00.7

S

21/12/00

66.2

38.1

38.9

93

6.6

7.2

691

00

02.8

SE

18/1/01

81.3

31.1

42.1

131

0.7

3.2

189

4.9

50

0.5

Calm

6/3/01

108

46.5

52.9

68

07.4

082

3.3

2.5

3.4

4.3

SE

7/3/01

87.2

43.7

35.3

107

1.7

14.3

7100

00

5.3

2.2

Calm

9/6/01

93.9

48.9

60.6

23

7.9

15.3

885

14.5

7.6

1.7

2.7

W

M. Price et al. / Atmospheric Environment 37 (2003) 4425–4434 4431

humidity is high throughout both the day and night. At

the Sunderland site the RH is high when atmospheric

particulate pollution is high whilst in the Albany study

the relative humidity was rising as particulate levels may

have been falling.

Table 3 shows that the meteorological conditions

found when there is a good agreement between the

monitors have tended to be dry with some sunshine

recorded. However, there are occasions when rainfall

has been recorded or the dew point reached and there is

no discrepancy between the monitors. The variation in

composition of atmospheric particulate matter will

determine the amounts of moisture adsorbed and this

could account for the variability in results obtained.

A thin film of water molecules coats many substances

under ambient conditions (Romakkaniemi et al., 2001)

and adsorption has been shown to occur at relative

humidities below the deliquescent relative humidity. Due

to its importance, the hygroscopic properties of aerosol

particles has been studied for a long time (Ebert et al.,

2002) including the work of Hanel (1976). A number of

studies have considered the adsorption of moisture by

atmospheric particulate matter including that by Hameri

et al. (2000), who have measured the hygroscopic

growth of ultrafine aerosol particles. However, in the

gravimetric measurement of atmospheric PM10

and PM2.5 it is assumed that once a particle is

deposited on a filter it remains inert even in the presence

of an air stream of high relative humidity containing a

range of gaseous pollutants. The work of Hameri et al.

(2000) and Romakkaniemi et al. (2001) would

suggest that this is unlikely to be the case and that

collected particles will continue to adsorb moisture at

relative humidities at or above the deliquescent relative

humidity for the sampled particles. In Sunderland,

particularly during the winter, the recorded ambient

relative humidity is above the deliquescent relative

humidity for particulate matter commonly sampled by

gravimetric monitors.

Although the gain of moisture by atmospheric

particulate matter associated with filters has been

illustrated by a number of researchers including the

work of Jarrett et al. (2001) its subsequent removal or

retention during filter conditioning has yet to be

determined. When the relative humidity decreases the

deliquescent RH will be reached, however, the

particle does not lose all of its associated moisture at

this or a decreased relative humidity (Seinfeld and

Pandis, 1998). In order to regain a solid state, salt nuclei

need to be formed and salt crystals must grow around

them. In the atmosphere this occurs at a relative

humidity significantly lower than the deliquescent

relative humidity. For salts such as ammonium nitrate

with a deliquescent relative humidity of 61.8% this may

well be below the relative humidity of 50% chosen for

filter conditioning.

ARTICLE IN PRESS

Table

3

Datesofgoodagreem

ent(o

5mg

m�3)betweenthePartisolandTEOM

monitors

Date

Partisol

TEOM

TEOM+SES

NO

xMin

temp

Maxtemp

Dew

point

RH

Rainfall

Rainfall

Sunshine

Meanwindspeed

Winddirection

(mgm

�3)

(mgm

�3)

(mgm

�3)

(ppb)

(�C)

(�C)

(�C)

(%)

(mm)

(h)

(h)

(ms�

1)

6/11/00

23.0

21.4

19.3

28

4.8

9.6

496

23.3

16

06.1

Calm

16/12/00

21.2

18.1

18.1

50

�1.0

4.2

�4

74

00

1.2

1.9

W

29/1/01

22.4

24.9

18.9

0.4

6.6

191

00

7.7

1.5

W

24/4/01

52.4

53.1

51.1

89

3.3

10.3

794

0.6

1.5

7.2

2.3

Calm

9/5/01

45.1

47.9

41.9

35

4.5

12.6

784

00

2.3

2.3

NE

31/5/01

20.8

20.6

18.3

20

9.4

17.4

866

2.3

0.4

12.6

5.3

W

3/6/01

15.2

16.8

14.0

12

6.0

14.7

358

00

6.8

3.8

N

1/8/01

26.7

24.9

22.5

35

12.0

24.5

10

64

00

12.3

1.9

W

14/8/01

21.6

22.5

20.9

29

18.8

24.7

17

70

0.7

0.4

10.1

3.9

SW

M. Price et al. / Atmospheric Environment 37 (2003) 4425–44344432

4. Conclusion

Particulate matter both PM10 and PM2.5 will comprise

a range of chemical species plus variable amounts of

particle bound water, where the amount of moisture will

be dependent on the composition of the atmospheric

aerosol and the relative humidity of the atmosphere.

This will vary both temporally and spatially.

The default operating temperature chosen for the

TEOM of 50�C; minimises the amounts of particle

bound moisture monitored. This has led to criticisms

that the TEOM under-records atmospheric particulate

matter due to the loss of semi-volatile materials such as

organics and nitrates. However, particle bound moisture

could be classified as a semi-volatile constituent of

atmospheric aerosols which is lost by the TEOM but

retained by gravimetric monitors. The results of the

monitoring campaign carried out in Sunderland, using a

TEOM fitted with a drier and operated at a lower

temperature, would suggest that at this site the gravi-

metric monitor is retaining a greater proportion of

atmospheric moisture than the TEOM monitor.

At this site it would therefore not be possible to apply

a factor to adjust the TEOM results to make them

comparable with the EU reference method, in that the

recorded differences are not consistent but governed by

aerosol composition and atmospheric relative humidity.

In addition application of a factor would unnecessarily

penalise local industry in that the increased amounts of

particulate matter recorded by the gravimetric monitor

could be due to atmospheric moisture and not an

increased anthropogenic input.

In any review of monitors and air quality standards

the role of particle bound moisture needs to be

evaluated.

References

Allen, G., Sioutas, C., Koutrakis, P., Reiss, R., Lurmann,

F.W., Roberts, P.T., 1997. Evaluation of the TEOM

method for measurement of ambient particulate mass in

urban areas. Journal of the Air and Waste Management

Association 47, 682–689.

APEG, 1999. Source apportionment of airborne particulate

matter in the United Kingdom. Airborne Particulates

Expert Group.

Ayers, G.P., Keywood, M.D., Gras, J.L., 1999. TEOM vs.

manual gravimetric methods for determination of PM2.5

aerosol mass concentrations. Atmospheric Environment 33,

3717–3721.

BSI, 1998. BS EN 12341:1998. Air quality—determination of

the PM10 fraction of suspended particulate matter—

reference method and field test procedure to demonstrate

reference equivalence of measurement methods. British

Standards Institute.

Chung, A., Chang, D.P.Y., Kleeman, J., Perry, K.D.,

Cahill, T.A., Dutcher, D., McDougall, E.M., Stroud, K.,

2001. Comparison of real-time instruments used to monitor

airborne particulate matter. Journal of the Air and Waste

Management Association 51, 109–120.

Cruz, C.N., Pandis, S.P., 2000. Deliquescence and hygroscopic

growth of mixed inorganic–organic aerosol. Environment

Science and Technology 34, 4313–4319.

Cyrys, J., Dietrich, G., Kreyling, W., Tuch, T., Heinrich, J.,

2001. PM2.5 measurements in ambient aerosol: comparison

between Harvard Impactor (HI) and the tapered element

oscillating microbalance (TEOM) system. The Science of

the Total Environment 278, 191–197.

Eatough, D.J., Long, R.W., Modey, W.K., Eatough, N.L.,

2003. Semi-volatile secondary organic aerosol in urban

atmospheres: meeting a measurement challenge. Atmo-

spheric Environment 37, 1277–1292.

Ebert, M., Inerle-Hof, M., Weinbruch, S., 2002. Environmental

scanning electron microscopy as a new technique to

determine the hygroscopic behaviour of individual aerosol

particles. Atmospheric Environment 36, 5909–5916.

Green, D., Fuller, G., Barratt, B., 2000. ‘Marylebone Road

(‘Supersite’), Annual Report, 199. http://www.seiph.umds.

ac.uk/detr/ss reports/reportar98.htm.

Green, D., Fuller, G., Barratt, B., 2001. Evaluation

of TEOM ‘correction factors’ for assessing the EU

Stage 1 limit values for PM10. Atmospheric Environment

35, 2589–2593.

Hameri, K., Vakeva, M., Hansson, H.C., Laaksonen, A., 2000.

Hygroscopic growth of ultrafine ammonium sulphate

aerosol measured using an ultrafine tandem differential

mobility analyzer. Journal of Geophysical Research-Atmo-

spheres 105, 22231–22242.

Hanel, G., 1976. The properties of atmospheric aerosol

particles as functions of the relative humidity at thermo-

dynamic equilibrium with the surrounding moist air.

Advances in Geophysics 19, 74–183.

Harrison, R.M., Deacon, A.R., Jones, M.R., 1997. Sources and

processes affecting concentration of PM10 and PM2.5

particulate matter in Birmingham (UK). Atmospheric

Environment 31 (24), 4103–4117.

Jarrett, R.P., Clark, N.N., Gilbert, M., Ramamurthy, R., 2001.

Evaluation and correction of moisture adsorption and

desorption from a tapered element oscillating microbalance.

Powder Technology 119 (2–3), 215–228.

Meyer, M., Lijek, J., Ono, D., 1992. Continuous PM10

measurements in a woodsmoke environment, PM10 stan-

dards and non-traditional particulate controls. In: Chow,

J.C., Ono, D.M. (Eds.), Journal of the Air and Waste

Management Association 21 (1), 24–38.

Meyer, M.B., Patashnick, H., Ambs, J.L., Rupprecht, E., 2000.

Development of a sample equilibration system for the

TEOM continuous PM monitor. Journal of the Air and

Waste Management Association 50, 1345–1349.

Mignacca, D., Stubbs, K., 1999. Effects of the equilibration

temperature on PM10 concentrations from the TEOM

method in the Lower Fraser Valley. Journal of the Air and

Waste Management Association 49 (10), 1250–1254.

Muir, D., 2000. The suitability of tapered element oscillating

microbalance (TEOMs) for PM10 monitoring in Europe.

The use of PM10 data as measured by TEOM for

ARTICLE IN PRESSM. Price et al. / Atmospheric Environment 37 (2003) 4425–4434 4433

compliance with the European Air Quality Standard.

Atmospheric Environment 34, 3209–3212.

Mukerjee, S., Shadwick, D.S., Bowser, J.J., Carmichael, L.Y.,

1999. Application of the dual fine particle sequential

sampler, a tapered element oscillating microbalance and

other air monitoring methods to assess transboundary

influences of PM2.5. Field Analytical Chemistry and

Technology 3 (3), 201–217.

QUARG, 1996. Airborne particulate matter in the United

Kingdom. Third report of the Quality of Urban Air Review

Group. May 1996, Department of the Environment. ISBN 0

0952077132.

Romakkaniemi, S., Hameri, K., Vakeva, M., Laaksonen, A.,

2001. Adsorption of water on 8–15nm NaCl and

(NH4)2SO4 aerosols measured using an ultrafine tandem

differential mobility analyzer. Journal of Physical Chemistry

A 105 (35), 8183–8188.

Salter, B., Parsons, L.F., 1999. Field trials of the TEOM and

partisol for PM10 monitoring in the St Austell china

clay area, Cornwall, UK. Atmospheric Environment 33,

2111–2114.

Seinfeld, J.H., Pandis, S.N., 1998. Atmospheric Chemistry and

Physics. From Air Pollution to Climate Change. Wiley

Interscience, New York.

Yang, H., Qianfeng, L., Jian Zhen, Y., 2003. Comparison of

two methods for the determination of water-soluble organic

carbon in atmospheric particles. Atmospheric Environment

37, 865–870.

ARTICLE IN PRESSM. Price et al. / Atmospheric Environment 37 (2003) 4425–44344434